Edisto, I'm realize you're a very intelligent man, and have offered some good testing advice. I appreciate that. Reading your background, it's understandable that my "experimenting" methods, or lack there of, bother you. This is what you do, you know how to do it well, and it bugs you to see someone screw it up. I understand that, I really do.
Doesn't bother me, and I didn't say you screwed it up. Everyone operates under constraints, and those constraints do not necessarily mean that an experiment cannot be done. It just means the results have to be interpreted with care, and that one should be extremely cautious about throwing numbers around.
However, I don't have the time, wood, or ambition to do testing to the level you're talking about. I only set out to get a good idea how the saw would perform at different levels of modification and compare that to a 346. In my own mind, I satisfied that.
I think you misunderstood my recommendations. You do have the resources. One simple change will add a lot to the comparability of future results. Instead of making X number of cuts with saw 1 followed by X number of cuts with saw 2, make the same number of cuts, but alternate the saws. That should reduce the variation due to inconsistency of the cant.
Then, to make the comparison, you can compare the averages.
To compare pre- and post mod, where you cannot alternate the cuts, you would need more cuts because you cannot alternate. In this case, you might not have enough wood to do enough cuts, especially if the wood is inconsistent. The more inconsistent the wood is, the larger the number of cuts you would have to make with each saw.
From your vantage point, my finding are completely dismisable, because they don't stand up to the rigor of a well designed scientific and repeatable test. But they told me what I wanted to know.
I never said that your findings are dismissable...I'm just saying you cannot give them too much weight. You had a lot of variability in the cant, but even accounting for that, the 261 is faster in the cut. That's a reasonable conclusion. It is not, however, reasonable to say that it is X% faster/stronger/whatever. Does that make sense? Estimates are meaningless without quantification of the error of that estimate. Quantifying the error requires a larger sample size.
It's all about putting the findings in the right context. Even if you had a lot of replication and remarkably consistent wood, or even if you had used a dyno to show the difference, it still is one 346 vs 1 261. You put a lot of effort into getting a single data point, and now we have to wait and see what the trend will be as other data points come to light.
Thus far, of all the comparisons made, the 261 came out on top.
You stated that my videos are unrepresentative of the actual results. The problem here is that I did not create the videos to be used as you did. Not only did I want to show the fastest cut times the saw was capable of, but I wanted to show how the saw ran under varying loads. Therefore you cannot use all cut times shown as the best cut times the saw is capable of. Again, they met my goals, but not yours as scientific test data.
Actually, you stated the videos were not representative, I just reiterated it. Just remember, comparisons should not be based on the fastest times, but on the average of all times that reflect a certain set of conditions.
Yes, I have lost my sense of humor. I understand your view point, but you have come across as simply trying to discredit everything I've shown here. While not perfect testing, I put a LOT of effert into this, and accomplished what I set out to do.
It is quite possible that I came across this way, but I took pains to not come across this way. Might be a combination of my own social retardation, and your persecution complex. You get ragged on a lot, I understand that, but don't use it as an excuse to ignore criticisms that you could learn from.
Thank you for your cordial reply. We're just on different pages here
There really is only one valid page, but no-one can force you to read it.